MEK2 Negatively Regulates Lipopolysaccharide-Mediated IL-1β Production through HIF-1α Expression

Mitogen-activated protein kinase (MAP2K, or MEK) 1 and 2 is a member of the dual specificity protein kinase family. In mammals, there are two distinct gene isoforms of Mek present, Mek1 and Mek2. Whereas the deletion of the Mek1 gene leads to embryonic lethality, interruption of Mek2 is compatible with life (1, 2). Both isoforms are considered to be directly upstream of ERKs (3). However, recent evidence suggests that each isoform has a unique biological role. For instance, MEK1 is capable of stimulating epidermal proliferation, and in fibroblasts it has a regulatory function in cell migration (2, 4). Furthermore, MEK1-deficient mice exhibit a lupus-like syndrome through deregulation of phosphatase and tensin homolog (PTEN) and protein kinase B (AKT) activation (5). The physiological role of MEK2 versus MEK1 in the innate immune system, especially in macrophages, is poorly understood (6, 7).

In contrast to the well-defined role of the MEK/ERK pathway in cell growth and cancer biology, the differential roles of MEK1 and MEK2 in response to TLR activation is poorly understood. TLR receptors are type I transmembrane proteins that mediate the recognition of pathogen-associated molecular patterns (8). The TLR family of receptors is composed of up to 10 members in humans and 12 in mice (9). Docking of LPS to TLR4 recruits the adaptor protein MyD88, which activates MAPKs, including ERK, JNK, and p38 kinase. TLR4 activation leads to phosphorylation of MEK1/2 and subsequent ERK1/2 activation. ERK1/2 activation has been proposed to play a major role in NF-κB activation, ROS and cytokine production especially IL-1β (10, 11). IL-1β production is tightly regulated through activation of several transcription factors.

The hypoxia-inducible factor (HIF)–1α belongs to the oxygen-sensitive transcription factors and is known as a transcriptional regulator for several inflammatory cytokines, including IL-1β and IL-6 (12, 13). In normoxic conditions cytosolic HIF-1α is hydroxylated by prolyl-hydroxylases (PHDs) on the α-subunit regulating targeted polyubiquitination and degradation via the von Hippel–Lindau (VHL)-dependent pathway (14). Mutations in pVHL and loss of its function may lead to HIF-1α accumulation and give rise to various cancers (15). In addition to pVHL loss of function, various conditions may lead to HIF-1α accumulation through VHL-independent pathways (16). Several mechanisms, including ferritin-mediated iron sequestration or activation of pathways including PI3 kinase, mTOR, ERK1/2, and GSK3β, have been proposed to regulate HIF-1α (17–20). It is well recognized that in response to TLR4 activation, HIF-1α protein escapes proteasomal degradation and dimerizes with HIF-1β, which facilitates its translocation to the nucleus (13, 21, 22). The exact LPS-mediated signaling leading to accumulation of HIF-1α and IL-1β production has not been fully elucidated. It has been shown that endotoxins can induce HIF-1α at the transcriptional level and increase its stability (12, 17, 23, 24).

We investigated the role of MEK2 in macrophages in response to LPS-mediated cytokine production applying a genetic approach. Using bone marrow–derived macrophages (BMDMs) derived from wild-type (WT), Mek1^d/d Sox2^Cre/+, and Mek2^−/− mice, we show that despite increased pVHL, MEK2-deficient BMDMs exhibit significantly higher HIF-1α levels at baseline and in response to LPS challenge. Higher HIF-1α levels in MEK2-deficient BMDMs was linked to a higher IL-1β production in response to LPS challenge. Furthermore, the abundance of HIF-1α and IL-β production was independent of ERK activation.

LPS (055-B5 ultrapure) was purchased from InvivoGen (San Diego, CA). Phospho-specific Abs against phospho-MEK1/2, ERK1/2, p38, and JNK as well as total ERK1/2, JNK, p38, MEK1, MEK2, VHL, and β-actin were purchased from Cell Signaling Technology (Beverly, MA). Glucose transporter 1 (Glut1) Ab was purchased from Thermo Fisher Scientific (Waltham, MA). IL-1 β Ab was purchased (R&D Systems). The HIF-1α Ab was purchased from Bioss (Woburn, MA). NLRP3 Ab was obtained from AdipoGen (San Diego, CA). HRP-conjugated anti-mouse and anti-rabbit IgG secondary Abs were purchased from Cell Signaling Technology, and HRP-conjugated anti-goat Ab was purchased from Santa Cruz Biotechnology (Santa Cruz, CA).

Animal studies were approved by the University Committees on Use and Care of Animals. WT, Mek1^d/d Sox2^Cre/+, and Mek2^−/− were generated onto the 129-cc background as previously described (1, 25). WT, Mek1^d/d Sox2^Cre/+, and Mek2^−/− mice were maintained at animal facilities at the Centre de Recherche Sur le Cancer de l'Université Laval (Quebec City, QC, Canada). Animal studies were approved by the corresponding Committee for Protection of Animals at Laval University on Use and Care of Animals. All experiments were performed in accordance with relevant guidelines and regulations.

BMDMs from mice were prepared as described previously (26). Briefly, femurs and tibias from 6- to 12-wk-old mice were dissected, and the bone marrow was flushed out. Macrophages were cultured with IMDM supplemented with 30% L929 supernatant containing glutamine, sodium pyruvate, 10% heat-inactivated fetal FBS, and antibiotics for 5–7 d.

RAW 264.7 cells were obtained from American Type Culture Collection and grown in DMEM with 10% FBS and 1% penicillin and streptomycin. DNA constructs were created as follows: cDNA encoding murine MEK2 was subcloned into the p3XFLAG-myc-CMV-24 vector containing a C-terminal Myc tag by PCR using specific primers. Transfection of RAW 264.7 cells was conducted using Lipofectamine LTX (Invitrogen) and a fixed amount of DNA (2.5 μg of DNA per well of six-well plate). Twenty-four hours after transfection, cells were treated with LPS and then harvested for immunoblotting.

BMDMs were seeded with a density of 2 × 10⁵ per well in six-well plates. Cells were transfected with either HIF-1α short interference RNA (siRNA) (L-040794-00-0005; Thermo Fisher Scientific) or with a scrambled siRNA pool (D-001810-10-05; Thermo Fisher Scientific) using Lipofectamine 2000 (Invitrogen). After 24 h, transfected cells were treated with LPS and then harvested for immunoblotting.

After the appropriate treatments, cells were washed with PBS and harvested in RIPA buffer (MilliporeSigma, MA) containing protease inhibitor and antiphosphatase mixtures, as previously described (27). Equal amounts of proteins (15 μg) were mixed with the same volume of 2× sample buffer, separated on 10% SDS-PAGE, and transferred to a polyvinylidene difluoride (PVDF) membrane (Bio-Rad, Hercules, CA) at 20 V for 1 h using a semidry transfer cell (Bio-Rad) as previously described (27). The PVDF membrane was blocked with 5% dry milk in TBST (TBS with 0.1% Tween 20), rinsed, and incubated with primary Ab overnight. The blots were washed and incubated with HRP-conjugated secondary anti-IgG Ab. Membranes were washed and immunoreactive bands were visualized using a chemiluminescent substrate (ECL Plus; GE Healthcare, Pittsburgh, PA). Images were captured on Hyblot CL film (Denville Scientific, Metuchen, NJ). OD analysis of signals was performed using ImageQuant software (version 5; Molecular Dynamics).

Murine TNF-α, IL-1β, IL-12, IL-10, and IL-6 cytokine levels in cell culture supernatants were measured using ELISA DuoKits (R&D Systems) as previously described (28).

The nitrite concentration in the culture media was used as a measure of NO production and was determined by measuring nitrite accumulation in the medium using Griess reagent (28).

Cytoplasmic and nuclear fractions were separated as described previously (27). Briefly, after treatment, the cells were resuspended in a hypotonic buffer (10 mm HEPES, pH 7.9, 0.5% Igepal, 2 mm MgCl₂, 10 mm KCl, 0.1 mm EDTA, 0.5 mm PMSF, 1.0 μg/ml leupeptin, and 1.0 μg/ml aprotinin) and incubated on ice for 10 min. After centrifugation at 14,000 × g for 1 min at 4°C, the supernatant (cytoplasmic) and the pellets (nuclear fraction) were collected.

BMDMs were lysed in RIPA buffer (MilliporeSigma) containing protease inhibitor and antiphosphatase mixtures, and the lysate was centrifuged to pellet the cell debris. The resulting supernatant was used for immunoprecipitation. First, the complex between immunoprecipitation matrix (IP-matrix; ImmunoCruz IP/WB Optima F, Santa Cruz Biotechnology) was formed by incubating HIF-1α Ab with IP-matrix overnight at 4°C with gentle rotation. After the incubation, the IP-matrix pellet bound to HIF-1α Ab was washed three times with cold PBS. After the wash, the IP-matrix bound to HIF-1α Ab was incubated with the lysate overnight at 4°C with gentle rotation. Immunoprecipitation pellet was washed four times with cold PBS with a brief centrifugation. Samples from the IP-pellets were resuspended in the loading buffer and subjected to gel electrophoresis.

Statistical analyses were performed using SPSS software, version 24.0 (SPSS; Chicago, IL). One-way ANOVA test and post hoc repeated measure comparisons (least significant difference) were performed to identify differences between groups. ELISA results were expressed as mean ± SEM. For all analyses, two-tailed p values <0.05 were considered significant.

Total RNA was extracted using TRIzol reagent (Life Technologies) and reverse transcribed using the Reverse Transcription System (Promega, Madison, WI). The primers (targeting IL-1β, ARNT, vascular endothelial growth factor [VEGF], HIF-1α, and a reference gene, GAPDH) were used to amplify the corresponding cDNA by using iQ SYBR Green Supermix (Life Technology). Quantitative analysis of mRNA expression was performed using the MX3000p instrument (Stratagene, La Jolla, CA). PCR amplification was performed in a total volume of 20 μl containing 2 μl of each cDNA preparation and 20 pg of primers (Life Technology). The PCR amplification protocol was performed as described previously (27). Relative mRNA levels were calculated after normalizing to GAPDH. Data were analyzed using ANOVA or two-tailed Student t test, and the results were expressed as relative fold change. The following primers were used in the PCRs: GAPDH, forward (5′-GTGAACGAGAAGGACTATAACCC-3′) and reverse (5′-GGCTGTGTACCAATGGACTG-3′); IL-1-β, forward (5′-CGCAGCAGCACATCAACAAGAGC-3′), and reverse (5′-TGTCCT CATCCTGGAAGGTCCACG-3′); HIF-1α forward (5′-TCAAGTCAGCAACGTGGAG-3′) and reverse (5′-TATCGAGGCTGTGTCGACG-3′);VEGF, forward (5′-GTAAGCTTGTACAAGATCCGCAGACG-3′) and reverse (5′-ATGGATCCGTATCAGTCTTTCCTGG-3′).

In macrophages, it is not known whether MEK isoforms play a differential role in LPS-mediated IL-1β production. Because there are no chemical inhibitors that specifically target MEK1 versus MEK2, we applied a genetic approach to further elucidate the differential role of MEK isoforms by using BMDMs derived from MEK1- and MEK2-deficient mice. Some studies suggested that MEK/ERK regulates TLR-mediated NO production (29, 30). NO production by macrophages plays an important role in innate immunity. To investigate the relative role of MEK1 and MEK2 in LPS-mediated NO production, murine BMDMs from WT and Mek1^d/dSox2^Cre and MEK2^−/− mice were challenged with LPS for 24 h, followed by measurement of nitrite in conditioned media as a relative indicator of NO production. As shown in Fig. 1A, the difference was short of statistical significance, but both MEK1- and MEK2-deficient BMDMs responded to LPS with lower production of NO compared with WT BMDMs. Next, we analyzed several other cytokines, including TNF-α, IL-10, IL-12, IL-6, and IL-1β. LPS challenge of all three strains of BMDMs led to the production of TNF-α (Fig. 1B). MEK2 deficiency had no significant effect on TNF-α production in response to LPS as compared with WT BMDMs. In contrast, MEK1 deficiency led to a significantly lower TNF-α production in response to LPS as compared with WT and MEK2^−/− (Fig. 1B). Similarly, LPS treatment resulted in a significantly lower amount of IL-10 production in MEK1- and MEK2-deficient cells as compared with WT macrophages (Fig. 1C). Consistent with our previous work, we observed that LPS challenge of MEK1-deficient BMDMs led to higher IL-12 production as compared with WT and MEK2-deficient cells (Fig. 1D). BMDMs from MEK2-deficient mice exhibited a higher IL-6 production in response to LPS (Fig. 1E). Interestingly, BMDMs lacking MEK2 responded to LPS challenge with a significant higher IL-1β production as compared with BMDMs from WT and Mek1^d/dSox2^Cre mice (Fig. 1F). It is known that MEK/ERK activation is a positive regulator of several cytokines (31, 32). Yet, strikingly, we found that MEK2 and MEK1 differentially modulate the response to TLR ligands in macrophages. Whereas MEK1 regulates IL-12, IL-10, and TNF-α, MEK2 regulates IL-1β and IL-6 production. MEK2 deficiency has minimal effect on TNF-α or NO production.

IL-1β is one of the most potent inflammatory cytokines that regulates host responses to infections (13, 33, 34). IL-1β is regulated through complex mechanisms at the transcriptional and posttranscriptional level. HIF-1α is an important transcription factor regulating IL-1β production in response to TLR4 stimulation (13). To determine if the enhanced IL-1β expression in response to LPS is regulated through a transcriptional mechanism, we evaluated IL-1β mRNA expression levels using quantitative RT-PCR (qRT-PCR). BMDMs from WT and MEK2-deficient mice were cultured side by side under equal conditions and stimulated with LPS for 1 h, and IL-1β mRNA expression was assessed. Fig. 2A shows the mean of relative gene expression for IL-1β normalized to GAPDH. MEK2-deficient BMDMs responded to LPS with a marked increase in IL-1β mRNA expression (2-fold) as compared with WT BMDMs. Because our results showed an increase in mature IL-1β in MEK2-deficient BMDMs in response to LPS (Fig. 1F), we tested whether unprocessed IL-1β (pro–IL-1β) is similarly increased. BMDMs derived from WT, MEK2^−/−, and Mek1^d/d Sox2^Cre/+ mice were challenged with LPS for 3 h. As shown in Fig. 2B, LPS stimulation led to pro–IL-1β expression in BMDMs derived from all three mice strains; however, BMDMs deficient in MEK2 exhibited a significantly higher level of pro–IL-1β as compared with WT. In contrast, we observed a lower pro–IL-1β expression in response to LPS in MEK1-deficient BMDMs. The NLRP3 protein levels in resting macrophages are thought to be insufficient for inflammasome activation. Microbial components such as TLR ligands or endogenous molecules induce NLRP3 expression through the activation of NF-κB (35). Because the activation of NLRP3 is important for mature IL-1β release and we have seen increased IL-1 β on multiple levels (mRNA, pro–IL-1β, and released IL-1β), we evaluated the effect of MEK1 and 2 deficiency on NLRP3 protein levels in response to LPS. WT BMDMs responded to LPS challenge after 60 min with NLRP3 induction. Surprisingly, we observed a more rapid and robust increase in NLRP3 levels starting at 30 min after LPS challenge in MEK2-deficient macrophages. Interestingly, MEK1-deficient macrophages exhibited minimal NLRP3 induction in response to LPS (Fig. 2D). Next, we investigated the effect of MEK2 deficiency on HIF-1α expression. We observed a significantly higher HIF-1α mRNA expression in MEK2-deficient BMDMs (6-fold) as compared with WT after LPS stimulation (Fig. 2F). HIF-1α heterodimerizes with the aryl hydrocarbon receptor nuclear translocator (ARNT–HIF-1β), which facilitates its translocation into the nucleus (36). In MEK2-deficient BMDMs, ARNT mRNA expression was significantly increased in response to LPS (Fig. 2G). Because HIF-1α transcriptional activity depends on HIF-1α protein stability, we assessed HIF-1α protein levels in BMDMs derived from WT and MEK2-deficient mice in response to LPS. LPS stimulation led to a time-dependent higher HIF-1α protein level in MEK2-deficient cells as compared with WT (Fig. 2H, 2I). These results suggest that MEK2 deficiency leads to a higher HIF-1α and IL-1β expression at both the mRNA and the protein levels in response to LPS challenge in murine BMDMs. Several studies have shown that HIF-1α accumulation largely depends on decreased degradation due to a decreased hydroxylation by PHDs in response to LPS. There are three PHD isoforms with distinct specificity for different prolyl hydroxylation sites within HIF-1α (37). Furthermore, all three PHD genes are under HIF-1α transcriptional control (38). We also assessed the mRNA expression for all three PHDs after LPS challenge. We observed that mRNA expressions of all three PHDs were significantly higher in MEK2-deficient BMDMs (data not shown). The augmentation of HIF-1α level in MEK2^−/− BMDMs was not due to lack of PHDs expression.

Next, we assessed the nuclear accumulation of HIF-1α in BMDMs derived from WT, MEK1-, and MEK2-deficient mice. BMDMs were cultured under normoxic conditions and nuclear and cytoplasmic protein fractions were assessed for HIF-1α expression at baseline and in response to LPS. The fractionated proteins were subjected to Western blot analysis assessing for HIF-1α and pVHL. HIF-1α levels were significantly increased in the nuclear fraction of MEK2-deficient BMDMs even in the absence of LPS challenge (Fig. 3A). We observed the presence of pVHL in the cytoplasmic fractions, which increased in response to LPS in all three strains of mice. Interestingly, pVHL showed the highest expression in MEK2-deficient BMDMs (Fig. 3A). Members of the p300/CBP family interact(s) with HIF-1α in a DNA-bound complex to activate HIF-1α–dependent transcription of target genes (39, 40). Next, we assessed p300/CBP protein interaction with HIF-1α from MEK2-deficient BMDMs. BMDMs from WT and MEK2-deficient BMDMs were cultured, and HIF-1α was immunoprecipitated with anti–HIF-1α Ab. The extract was immunoblotted using HIF-1α (Fig. 3B) and p300/CBP Abs (Fig. 3C). As shown in Fig. 3C, there is significantly higher p300/CBP in HIF-1α–precipitated protein from MEK2-deficient BMDMs as compared with WT. These data indicate that in MEK2-deficient BMDMs, transcriptionally active HIF-1α accumulates in the nuclear fraction and recruits the transcriptional coactivator p300/CBP. Glucose is the primary metabolic substrate of macrophages in response to endotoxin stimulation. In response to endotoxin, Glut1 expression rapidly increases and regulates metabolic reprogramming to drive proinflammatory cytokine production, including IL-1β (41, 42). Additionally, HIF-1 α and the RAS/MEK pathway are known to regulate Glut1 expression (43). We assessed Glut1 expression in response to LPS in BMDMs derived from WT, MEK2^−/−, and Mek1^d/d Sox2^Cre/+ mice. Cell lysates were subjected to immunoblotting using an Ab against Glut1, and equal loading was confirmed using β-actin Ab. As shown in Fig. 3D, 3E, BMDMs from all three mice strains responded to LPS with increased Glut1 expression at 6 h. Both MEK1- and MEK2-deficient BMDM macrophages exhibit higher Glut1 at baseline. In response to LPS, MEK1-deficient BMDMs did not augment Glut1 expression, whereas MEK2-deficient BMDMs exhibit higher Glut1 at baseline, which was further augmented in response to LPS. These data indicate that MEK2^−/− BMDMs exhibit significantly higher Glut1 and IL-1β, both of which are transcriptionally controlled by HIF-1α. In addition to Glut1 and IL-1β, HIF-1α regulates VEGF expression (44). We determined the VEGF gene expression in BMDMs derived from WT, MEK1-, and MEK2-deficient mice. Fig. 3F shows a significantly higher expression for VEGF in MEK2-deficient BMDMs.

Because HIF-1α is a transcription factor regulating both IL-1β and Glut1, we investigate whether targeted downregulation of HIF-1α via siRNA can modify IL-1β level in response to LPS in MEK2-deficient macrophages. MEK 2^−/− BMDMs were transiently transfected with either non-sense vector (scrambled siRNA) or vector encoding HIF-1α siRNA. After 24 h transfection, cells were challenged with LPS. We observed a significant decrease in HIF-1α expression with targeted HIF-1α siRNA and a 80% decrease in pro–IL-1β in response to LPS, as compared with nontargeted siRNA transfected cells (Fig. 4A–C). Similarly, when we measured released IL-1β in conditioned medium of transfected cells after 24 h LPS challenge, we observed a significant reduction in IL-1β production (Fig. 4D). These data indicate that HIF-1α is an important transcription factor for IL-1β production in response to LPS in MEK2-deficient macrophages.

To address the relative contribution of MEK2 in LPS-mediated MAPK activation, we first determined the level of MEK1 and MEK2 protein expression by Western blot analysis in BMDMs derived from MEK2^−/− mice. As shown in Fig. 5A, MEK2-deficient BMDMs show no expression of MEK2 protein with comparable MEK1 protein levels in either WT or MEK2-deficient macrophages. To further investigate the contribution of MEK2 in LPS-mediated MEK1/2 phosphorylation, BMDMs from WT and MEK2^−/− mice were challenged with LPS for different time periods. LPS treatment of WT BMDMs led to phosphorylation of MEK1/2 (Ser217/221), which persisted up to 3 h (Fig. 5B). MEK2^−/− BMDMs showed a lower phosphorylation of MEK1/2 compared with WT BMDMs (Fig. 5B). The phosphorylation seen in MEK2-deficient BMDMs is mainly due to the presence of MEK1, as the Ab recognizes the phosphorylated forms of both MEK1 and MEK2. Furthermore, we assessed the role of MEK2 in the activation of other MAP kinases (ERK, p38, and JNK). Cell lysates were subjected to immunoblotting using Abs against the phosphorylated forms of ERK (Thr202/Tyr204), p38 (Thr180/Tyr182), and SAPK/JNK (Thr183/Tyr185). As expected, LPS stimulation led to ERK1/2 phosphorylation in WT BMDMs and in MEK2-deficient cells. This phosphorylation peaked at 30 min and persisted for up to 3 h (Fig. 5C). Densitometric analysis is shown in Fig. 5D. LPS treatment led to a rapid and similar pattern of p38 phosphorylation in WT, in both MEK1- and MEK2-deficient BMDMs (Fig. 5G). Densitometric analysis is shown in Fig. 5H. We also detected JNK phosphorylation. As shown in Fig. 5E WT, MEK1-, and MEK2-deficient BMDMs responded to LPS challenge with similar JNK phosphorylation. Densitometric analysis is shown in Fig. 5F. Taken together, our results suggest that MEK2 is dispensable in ERK, p38, and JNK phosphorylation in response to LPS stimulation of murine BMDMs.

Our data are consistent with the hypothesis that MEK2 negatively regulates IL-1β independent of ERK activation. Hence, we speculated that overexpression of MEK2 should prevent the production of IL-1β. To this end, we transfected RAW264.7 cells with CMV-MEK2 for 24 h. As shown in Fig. 6A, CMV-MEK2–transfected cells showed a significant increase in the expression of MEK2 protein compared with nontransfected (NT) cells. To further investigate the contribution of MEK2 in LPS-mediated MEK1/2 phosphorylation, cells from NT and CMV-MEK2 were cultured in the presence of LPS for 30 min. LPS treatment led to a 2-fold increase in MEK1/2 phosphorylation in CMV-MEK2–transfected RAW264.7 cells as compared with NT cells (Fig. 6A). Densitometric analysis is shown in Fig. 6B. To elucidate the role of MEK2 overexpression in regulating ERK activation in response to LPS, we challenged NT or CMV-MEK2–transfected RAW264.7 cells with LPS for 30 min and assessed for ERK phosphorylation. As shown in Fig. 6C, LPS stimulation led to ERK phosphorylation in NT RAW264.7 cells. Interestingly, ERK phosphorylation was not significantly changed in MEK2-overexpressed cells (Fig. 6C). Densitometric analysis is shown in Fig. 6D. Furthermore, we tested whether overexpression of MEK2 could decrease the level of pro–IL-1β. Cell lysates were subjected to immunoblotting using Abs against pro–IL-1β. As shown in Fig. 6E, as expected, LPS stimulation led to a high pro–IL-1β expression in NT RAW264.7 cells, whereas RAW264.7 cells overexpressing CMV-MEK2 responded to LPS with a significant lower pro–IL-1β level as compared with NT cells (Fig. 6E). Densitometric analysis is shown in Fig. 6F. These results suggest that MEK2 overexpression leads to a lower expression of IL-1β in response to LPS challenge with no significant effect on ERK phosphorylation.

Because MEK1 and MEK2 structurally similar, they are thought to be functionally redundant. However, recently Catalanotti et al. (45) have shown that MEK1 and MEK2 build a heterodimer complex containing a negative feedback loop. This raises the question of whether there is a role for a disrupted negative feedback loop through a lack of MEK1-MEK2 heterodimerization in MEK2-deficient BMDMs. To test this hypothesis, we generated mutants with three Mek allele deleted (triple mutants). Mice with only one Mek1 left (Mek1^f/+ Mek2^−/− Vav iCre⁺) are named 1M1L, and triple mutant with only one Mek2 left (Mek1^f/f Mek2^+/− Vav iCre⁺) are named 1M2L. We then investigated the role of Mek1 and Mek2 double and triple mutation on ERK phosphorylation, HIF-1α expression, and IL-1β production. BMDMs from WT, MEK2^−/−, 1M1L, and 1M2L were isolated and challenged with LPS. ERK1/2 phosphorylation in response to LPS in WT BMDMs or Mek2 double mutant mice were comparable, whereas 1M1L mutant exhibited a decreased level of phosphorylated ERK1/2 (Fig. 7A, 7B). BMDMs from triple mutant macrophages lacking two MEK1 (1M2L) responded with a significant diminished ERK phosphorylation to LPS challenge. Fig. 7A lower panel shows the expression for MEK1 and MEK2 proteins. The mean densitometric analysis is shown in Fig. 7B. Next, we determined the effect of these mutants on HIF-1α expression by Western blot analysis. WT BMDMs responded to LPS with a minimal expression of HIF-1α, whereas the absence of MEK2 led to increased HIF-1α expression in response to LPS challenge (Fig. 7C). BMDMs from the Mek2 double or 1M1L mutant mice exhibited a significantly increased level of HIF-1α protein expression in response to LPS challenge. BMDMs from triple mutant macrophages lacking two MEK1 (1M2L) exhibited higher baseline HIF-1α expression but no further increase in response to LPS (Fig. 7C). Mean densitometric analysis is shown in Fig. 7D. Next, we assessed IL-1β production in response to LPS in Mek1 and Mek2 double and triple mutations. As shown in Fig. 7E, the highest amount of IL-1β production was observed in macrophages derived from mice lacking MEK2 in response to LPS challenge. Importantly, introduction of one allele of MEK2 (1M2L) (Mek1^f/f Mek2^+/− Vav iCre⁺) led to a significant decrease of IL-1β production in response to LPS. These data suggest that rather a complete deletion of the MEK2 allele leads to higher IL-1β production and not the absence of MEK1, as the presence of MEK1 (1M1L) could not completely restore ERK phosphorylation and decreased IL-1β production. Fig. 7F demonstrates the relationship of MEK2 protein expression of BMDMs and IL-1β production in response to LPS.

Despite structural similarity between MEK1 and MEK2, evidence indicates that they are not functionally redundant. Several studies using genetic dissection of MEK1 and MEK2 in mouse models has suggested nonoverlapping functions in disease development (5, 45). MEK1 and MEK2 show high homology in their kinase domains, whereas their N termini and their proline-rich domains show only 40% identity, which provides an opportunity for the two isoforms to interact and partner differently with scaffolding proteins and their activators and substrates (46, 47). Although the role of MEKs is well studied in developmental and cancer biology, the specific role of MEK2 versus MEK1 in innate immunity is not well understood. Most previous reports used chemical inhibitors (mostly U0126 or PD98059) to study the role of MEK/ERK in cancer biology or in TLR-mediated activation (43, 48). However, these inhibitors are nonspecific for MEK1 versus MEK2 and have off-target effects on several other kinases (49). In most studies, physiological consequences of MEK1/2 activation are judged by the effect on the activation of their best-known substrate ERK1/2. Our investigation is the first study, to our knowledge, to describe a unique role of MEK2 in HIF-1α expression and IL-1β production after LPS challenge that is independent of activation of ERK, p38, and JNK in murine macrophages. Interestingly, MEK2^−/− macrophages exhibited early induction of NLRP3 in response to LPS associated with increased release of mature IL-1β. Furthermore, BMDMs derived from MEK2-deficient mice challenged with LPS exhibit a preserved ERK activation but significantly higher HIF-1α expression at both the mRNA and protei2n levels. In contrast, we observed a lower level of NLRP3 induction, reduced pro–IL-1β and mature IL-1β production, and a lower HIF-1 α level in response to LPS challenge in MEK1-deficient BMDMs.

HIF-1α is recognized as the master regulator of the hypoxic response, which modulates transcription of numerous genes (50). HIF-1α expression is critical for the metabolic switch during inflammation and in cancer, as its induction regulates glucose uptake and the expression of glycolytic enzymes (51). In cancer and during inflammation, aerobic glycolysis (Warburg effect) plays an important role in the maintenance of cellular energy supply, as the aerobic ATP production through the tricyclic acid cycle is suppressed (48, 52). This metabolic switch from oxidative phosphorylation to aerobic glycolysis is grossly coordinated by HIF-1α expression (51). Glucose is a major source of energy during inflammation and is transported across the plasma membrane facilitated by glucose transporters. Among these glucose transporters, Glut1 regulates the enhanced glucose uptake in macrophages as well as in cancer cells. Its expression is rapidly induced in response to LPS (53, 54). Glut1 downregulation decreases inflammatory response, especially IL-1β production (55, 56). Various kinases including GSK3/TSC/mTOR, PI3K/AKT, and RAS/MEK pathways have been proposed to regulate Glut1 expression (57). Surprisingly, in both MEK1- and MEK2-deficient macrophages, we found increased Glut1 levels at baseline as compared with WT. MEK2 deficiency exhibited higher Glut1 levels, which significantly increased in response to LPS.

Although HIF-1α accumulation was originally identified under hypoxia, as its name implies, now a large body of literature indicates that HIF-1α is strongly activated by TLRs, IFN-α, EGF and other cytokine receptors under well oxygenated conditions (12, 58, 59). There are several mechanisms underlying HIF-1α accumulation in response to LPS. TLR4 mediated induction of HIF-1α reflects a combination of increased HIF-1α transcription and decreased HIF-1α degradation (12). In this study, we show that HIF-1α transcripts and protein expression were significantly increased in responses to LPS in MEK2-deficient macrophages as compared with WT. Additionally, ARNT (HIF-1β), the binding partner of HIF-1α, was significantly increased in MEK2-deficient macrophages. Proline hydroxylation is important step in HIF-1α destabilization (37). In our study, augmented HIF-1α expression was not due to diminished expression of PHDs, as in MEK2-deficient BMDMs the expression for all three PHDs was higher in response to LPS. HIF-1α transcriptional activity is dependent on nuclear translocation (60). We observed that HIF-1α rapidly accumulates in the nuclear fraction in MEK2-deficient macrophages as compared with WT macrophages.

The molecular basis of how TLR4 mediates activation of RAS-GTPase upstream to MEK has not been precisely elucidated. Interaction of LPS or other lipid-based mediators (e.g., lipid A) leads to association of TLR4 and MD2, which recruits several adaptor proteins (61). This is thought to lead to the subsequent activation of receptor tyrosine kinases (RTKs) and the classical RAS–RAF–MEK1/2 cascade (62–64). Additionally, TLR4 ligation can activate ERK1/2 through alternative pathways independent of RAS–RAF–MEK (65, 66). For instance, LPS can activate MEK/ERK through PKC ζ (67). Recently, we have shown that the MEK1 isoform plays a critical role in ERK1/2 activation in response to LPS (7). Despite intact MEK2, MEK1-deficient macrophages responded only minimally to LPS with ERK activation (7), whereas MEK1-deficient macrophages responded to LPS challenge with enhanced STAT4 phosphorylation and a heightened IL-12 production but decreased IL-10 production. The simultaneous challenge of TLR4 and retinoic acid or rIL-10 restored ERK1/2 phosphorylation (7). This suggests that MEK2-mediated ERK1/2 activation requires additional signals (retinoic acid or IL-10) besides TLR receptor activation (7). Our current work shows that MEK2^−/− BMDMs respond to LPS stimulation similar to MEK1-deficient macrophages with lower IL-10; in contrast, IL-12 production in MEK2^−/− macrophages was significantly lower as compared with MEK1-deficient macrophages. Furthermore, we observed in MEK2-deficient macrophages strikingly higher production of IL-1β production in response to LPS challenge. Furthermore, overexpression of MEK2 in RAW264.7 cells decreased IL-1β production in response to LPS, whereas MEK2 overexpression did not modulate ERK activation (Fig. 6). Most interestingly, the IL-1β–producing phenotype of BMDMs from triple Mek knockout mice (lacking Mek1 but having one allele of Mek2) was rescued by introducing one allele of the Mek2 gene (Fig. 7). BMDMs of these mice challenged with LPS showed preserved ERK activation but a lower HIF-1α expression as compared with MEK2-deficient macrophages. These data indicate that the level of MEK2 regulates IL-1β production in response to LPS independent of ERK activation. Our previous and current work suggest a differential role of MEK1 and 2 in the regulation of cytokine networks in response to TLR4 activation. This observation is consistent with the differential role of the MEK isoforms in cancer cell biology (46), in which HIF-1α is thought to be central in controlling neovascularization, survival, and tumor spread in various cancers. Cancer-associated enhanced HIF-1α expression is predominantly attributed to oxygen limitation in the tumor microenvironment (68). Because the deregulation of Ras/Raf/MEK/ERK signaling cascades and HIF signaling plays a crucial central role in tumor growth and spread (69, 70) and our data indicate that MEK2 plays an important role in HIF signaling, it would be interesting to further investigate the role of MEK2 in cancers.

Our data suggest an exquisite role of MEK2 as negative regulator of HIF-1α and Glut1 expression as well as IL-1β production. Yet it remains unclear how the lack of MEK2 would influence the complex interplay of these pathways (71, 72). Further studies need to determine whether MEK2 directly or indirectly regulates HIF-1α expression. In summary, our data indicate that there is a unique and indispensable role for MEK2 in regulating HIF-1α and thereby modulating IL-1β in response to LPS challenge that cannot be substituted by MEK1. Thus, modulation of MEK2 may represent a novel therapeutic target in HIF-1α–mediated inflammation as well as cancer.

The authors have no financial conflicts of interest.